1 Executive summary


National deficit of around 58k admissions for hip replacements between the onset of national lockdown (26th March 2020) and the end of 2020. Proportionally affecting the North West and South West NHSE regions most, where inpatient admissions during 2020 were over 50% less than would be expected. Lowest levels of disruption in London (approx. 30%) perhaps linked to differing approach to recovery where clinical hubs were set up to provide speciality-specific care.


Admission deficit in planned care:

  • Elderly groups more disrupted (approx. 50% less admissions). Disruption reducing with age, and presumably linked to reducing risk perceptions.
  • Admission deficit/disruption proportionally highest in white British and other white ethnic sub-groups.
  • Small but significant deprivation gradient seen in disruption to care, highest in most deprived, lowest in most affluent. This relationship is not consistent across regions; extenuated in London (30% care disrupted in IMD quintile 1, approx. 17% disrupted in IMD quintile 5), while the South East and South West display inverse deprivation relationships.


Admission deficit in unplanned care:

  • National deficit closely matches lockdown timeline
  • Unclear how to interpret a reduction in ED attendances – indicative of more effective planned care or reduced demand due to decreased mobility/activity.
  • Highest proportional deficit in London, in comparison to lowest in London for planned admissions.
  • Deprivation gradient seen in unplanned and planned admission disruption.


COVID-19 impact on quality of life


As of January 2021, the increased waiting for hip replacements resulted in a loss of approx 895 QALYs. Converting this figure to a monetary value, at £20k per QALY and discounting by 3.5% per year, values the total lifetime loss at £223 million. In addition, the disutility of increased waiting can be estimated at £27 million.

2 Background

The COVID-19 pandemic has had hugely disruptive impacts on the provision and delivery of emergency and elective care. Such effects have been felt differently across the country and have been moderated by disease progression and changing lockdown measures. Underlying health and social inequities across the population have been exposed. Pre-existing disease rates, barriers to healthcare and economic instability have aligned to present disproportionate threats to health for certain demographics and geographical regions.

2.1 Hip replacements

GP referrals for orthopaedic elective care have seen one of the largest reductions during the pandemic adding to the already troubling waiting times for consultant led referrals [1]. Increased waiting times may lead to increases in emergency admissions for untreated hip pain and mobility issues leading to trips, falls and acute presentations at A&E . Emergency care systems may have increased thresholds by which patients are encouraged to attend; instead, patients were advised to seek virtual guidance or provide self-care where appropriate . Redirection of emergency care may have delayed attendance to A&E to the point of crisis, increasing the severity of such attendances, calling for more invasive treatment when seen and increasing costly disability during waiting periods [2]. The shift to virtual management of chronic hip pain and mobility issues has been vast but quality and uptake inevitably vary by provider and patient population, respectively. Virtual care brings potential for widening inequities when we consider differing economic accessibility, IT knowledge and healthcare literacy across age ranges and cultural groups.

3 Methods

3.1 Project scope

The strategy unit proposes to quantify the changes during the COVD-19 pandemic in hip replacement activity, including pre- and post-operative care, according to selected demographic, socio-economic and geographic population groups.

3.2 Data collection

3.2.1 Inpatient care

Inpatient care was sub-divided into either ‘elective’ or ‘unplanned’ admissions based on the spell admission method. OPCS procedure codes were used to identify admissions with any of the following procedures:

Primary replacements:

  • Total Hip replacement
  • Total Prosthetic Replacement Head Femur
  • Hybrid Prosthetic Hip
  • Other Hip Replacement Primary
  • Resurfacing Arthroplasty Joint Primary

Revisions:

  • Total Hip Replacement Revision
  • Total Prosthetic Replacement Head Femur Revision
  • Hybrid Prosthetic Hip Replacement Revision
  • Other Hip Replacement Revision
  • Resurfacing Arthroplasty Joint Revision

3.2.2 Outpatient appointments

Appointments including pre- or post-operative reviews relating to the above procedures and/or including any of the following diagnosis codes:

  • S700: Contusion of hip
  • S730: Dislocation of hip
  • S731: Sprain and strain of hip
  • S760: Injury of muscle and tendon of hip
  • S770: Crushing injury of hip
  • S772: Crushing injury of hip with thigh
  • S797: Multiple injuries of hip and thigh
  • S798: Other specified injuries of hip and thigh
  • S799: Unspecified injury of hip and thigh

3.2.3 Emergency department attendances

Unplanned attendances at emergency care settings for the following injuries/complaints were identified using SNOMED-CT codes:

  • Femur fracture (shaft and NOF)
  • Hip fractures (pubis, acetabulum, ilium, ischium)
  • Hip sprains
  • Hip dislocation
  • Open wounds to the hip (lacerations, abrasions and open fractures)
  • Hip pain (these are “findings” as opposed to “disorders”)

3.3 Analysis plan

Areas of interest include

  • What were the cohort characteristics of the patients who were prioritised (and inversely, those postponed) at the peak of the lockdown?
    • Does this appear to be clinically driven?
  • What were the cohort characteristics of the patients seeking emergency care?
  • How did the pandemic impact activity by care setting, disease type, demographic and geography?
  • What were the effects of delayed hip procedures on quality of life?

4 COVID-19 impact on activity

4.1 National time series

Fig. 1.1


The effect of COVID-19 on inpatient elective hip replacements is clear. The daily count of admissions fluctuated by weekday and weekend and there were regular reductions around Christmas and New Year Eve. A large drop is seen with the implementation of a national lockdown on 26th March 2020. Admission rates recovered and rose until September 30th 2020, after which daily admissions dropped gradually during the second national lockdown and the holiday period.



Fig. 1.2
COVID-19 impacted care settings differently; planned inpatient and outpatient care was canceled in line with national lockdown so trends follow similar patterns, while emergency department attendances saw less of a reduction but patients care seeking behaviour were clearly impacted to an extent.


4.2 Cohort characteristics

4.2.1 Age

Fig. 2.1

Fig. 2.2.1 & 2.2.2

4.2.2 Deprivation

Fig. 2.3

Fig. 2.4.1 & 2.4.2

Fig. 2.5

4.2.3 Ethnicity

Fig. 2.7

Fig. 2.8.1 & 2.8.2

Fig. 2.9

4.3 Deficit in planned admission

4.3.1 National and regional

Fig. 3.1
By calculating the difference between the 2020 trend and the 2018-19 average weekly number of admissions we can present the admission deficit. So far, there have been 58180 admissions less than we would expect for elective inpatient hip replacement admissions. The backlog of admissions is likely to last until after the deficit is zero.


Fig. 3.2
The initial effect of covid-19 and national lockdown brought elective inpatient admissions for hip replacements down to near zero in all areas. As such, high admission deficits are seen where the admission volume was high pre-pandemic and volumes did not return to the levels seen in previous years through the ‘Recovery’ period; specifically Midlands and North East and Yorkshire.


Fig. 3.3
Another influential factor in determining the volume of a region’s admission deficit is the magnitude of the second lockdown. Again the Midlands and North East and Yorkshire experienced particularly disruptive second lockdowns, as seen above.


Fig. 3.4
‘Deficit proportion’ is calculated below by dividing the region’s admission deficit by the average annual number of admissions seen between 2018 and 2019, as a proxy for the number of admissions we would have expected to see in the absence of the pandemic.
The above chart demonstrates that while the most patients were delayed or cancelled in the Midlands and North East and Yorkshire, the largest proportional effect of covid-19 on elective hip replacement admissions was seen in the North West; where over half of the expected activity was disrupted. Conversely, London and the South East achieved the lowest admission deficit proportion with only 30-40% of the expected number of admissions delayed or cancelled.


4.3.2 Demographic deficits

4.3.2.1 Age


Fig. 3.5
Though a spike in the proportion of admissions attributed to 80-99 year old’s was seen at the onset of national lockdown, the overall deficit demonstrates high levels of disruption in this group with over half of the expected admission volume for this group either delayed or cancelled. This effect is seen to deminish as you move down the age ranges, where the least disruption is seen in age ranges considered to have the lowest risk of covid-19 morbidity and mortality.


4.3.2.2 Ethnicity


Fig. 3.6
Ethnicities classified broadly as ‘White’ are the three most delayed or cancelled ethnic groups; ‘British’, ‘Irish’ and ‘Any other white background’. A large effect was also seen in the ‘Not Stated’ group, suggesting around 35% of the expected activity for this group was disrupted. Mixed effects were seen elsewhere, making conclusions by broad ethnic groups difficult.


Fig. 3.7
The respective representation of ethnic groups was largely the same between 2018-19 and 2020 other than increases seen in ‘Unknown’ ethnicities and relative decreases in the British group. Changes seen here may be linked by data collection practices during high stress periods, in that a large proportion of the increase in unknown may be coming from reductions in British patients.


Fig 3.8
Trends in the changing proportions of hip replacement procedures performed by ethnicity at the national level are echoed across the country. Reducing proportions of white hip replacement patients and increasing proportions of patients of ‘unknown’ ethnicity vary in terms of magnitude regionally but are directionally consistent.


4.3.2.3 Deprivation


Fig. 3.9
Hip replacement admissions for the most deprived 20% of the population were significantly more disrupted than each of the more affluent IMD quintiles.


Fig. 3.10
A deprivation gradient exists when comparing the changing proportion of elective hip replacement procedures between the 2018-19 average and those admitted during 2020. The proportion is seen to increase in more affluent groups and decrease is the more deprived members of the population. Where initially we saw a small increase in the proportion of admissions attributed to IMD quintile 1 (the most deprived 20% of the population) during the first lockdown, we then saw a generally decreasing trend for this group through the ‘recovery’ and ‘second lockdown’ periods.


Fig. 3.11
The patterns we see in admission deficit by deprivation quintile vary across the country. London demonstrates clear inequity with deprived people having their procedures delayed or cancelled much more than affluent people. It may also be the case that more deprived groups weren’t demanding care as much as the affluent were, potentially due to lack of sick leave/pay or more hesitancy to attend hospital during the pandemic; similar gradients can be seen in the North East and Yorkshire and the North West regions respectively. Conversely it seems that the case of more affluent populations were most disrupted by the pandemic in the South East and South West respectively.



Fig. 3.12
Deprivation gradients are present in Asian and White groups; to varying degrees, the more deprived members of these ethnic groups were more disrupted by the pandemic.


Fig. 3.13
Deprivation inequities in Asian populations seem to be driven by Pakistani and Indian subgroups. Where negative deficit proportions are seen, we can deduce excess admissions occurred; more admissions in this subgroup in 2020 than the 2018-19 average. This is the case with Bangladeshi patients however the magnitude of the excess is very small and therefore has little influence on the aggregated results for the Asian population.



Fig. 3.14
Increases in proportion of those receiving hip replacements are seen in all ethnicities other than the White population. The magnitude of the increase is much higher in the ‘Not known’ category than other non-White groups. The deprivation profile of the ‘Not known’ ethnicity appears to inversely match that of the White population supporting the hypothesis that increases in ‘Not known’ are primarily from reductions in the White patients.


Fig. 3.15
Inequalities in hip replacement admission deficit appear to be most accentuated in the 20-39 year old age range. Here we can see the most deprived members of the population are disproportionately disrupted while there is less of a deprivation gradient in the 60+ population, where the most hip replacements are performed.


4.4 Deficit in unplanned admissions

4.4.1 National and regional

Fig. 4.1
When viewing the trend in emergency, unplanned admissions for hip replacements, we see regular winter spikes in December 2018 and 2019. We also see a small reduction from the normal activity level around the first national lockdown with a gradual ‘recovery’


Fig. 4.2
Outside of national lockdown periods, we see very similar admission rates for unplanned hip replacement procedures. Factors that may be influential here are a) reduced population mobility (similar to trends seen in RTA’s?), b) changing ED thresholds for admitting patients who attend with hip injuries and/or c) reduced care seeking behaviour.


Fig. 4.3
The cumulative deficit trend above shows that while admissions rates similar to the previous years returned during the ‘recovery’ period, the deficit will remain until rates surpass the comparative trend.


Fig. 4.4




Fig. 4.4.2 (alternative plot)
…


Fig 4.5
There are clear and significant differences in disruption to emergency hip replacements across the country; it is unclear whether differences here are due to region differences in maintaining planned procedures rates throughout the pandemic or changing emergency department thresholds to admit patients.


Fig. 4.6
By comparing each region’s planned and unplanned admission deficit proportion we see two main outliers: London with a high unplanned deficit alongside a low planned deficit, while the South East has low planned and unplanned deficits. Admission rates in the South East have remained the most consistent with previous years in terms of planned and unplanned hip replacements. It is unclear from this analysis how the two rates interact and co-depend.

4.4.2 Demographic deficits

4.4.2.0.1 Deprivation


Fig. 4.7
We can see that the more affluent members of the population had the least disruptions to the unplanned admission rate. Attributing that trend to policy or behaviour is difficult. Were the more affluent groups more able to navigate the care system to demand care in spite of the pandemic? Were more deprived groups more apprehensive to seek care for fear of Covid-19?

Fig. 4.8
Viewing changes in planned admission deficits alongside that of unplanned care might identify reductions in one admission type and subsequent increases in the other; however when viewing the population by deprivation quintile, we see the more affluent people experienced the least disruption to planned and unplanned hip replacement care.

4.4.2.0.2 Ethnicity


Fig. 4.9
Broad confidence intervals suggest drawing conclusions from ethnicity groups with small numbers may be ill-advised. We can say with confidence that the emergency hip care for Pakistani, Indian and Caribbean members of the population was disproportionately disrupted, in comparison to British and other white groups.

Fig. 4.10
When comparing proportional admission deficit for planned and unplanned care, it is clear the Asian sub-groups are experiencing higher than average levels of disruption in planned and unplanned admissions for hip replacements and related care; the same could be said for white British populations.

5 Quality of life effects

As of January 2021, the increased waiting for hip replacements resulted in a loss of approx 895 QALYs. Converting this figure to a monetary value, at £20k per QALY and discounting by 3.5% per year, values the total lifetime loss at £223 million. In addition, the disutility of increased waiting can be estimated at £27 million.

These figures will continue to rise until the deficit is reduced, that is we recover all the activity lost due to covid-19. Details below:

  • Literature derived [2] QALY impact of 1 week delay in hip replacement surgery: 0.00062
  • Total admission deficit: 58180 admissions
  • Cumulative additional weeks waiting for hip replacement: 1444251 weeks
  • QALYs lost by those waiting for hip procedures: 895.4 QALYS
  • Monetary value of QALYs lost by those waiting (year 1): £17908712
  • Monetary value of lifetime lost of QALYs: £223206439
  • Disutility of waiting for hip procedures: £27440769




6 Adittional analysis

6.1 London-style recovery modelling


How would the national deficit in elective hip replacements vary if all regions had recovered like London?


We modelled an alternative admission trend for each region to match the recovery profile seen in London; that is, near-zero admission rates throughout the first lockdown, and a steady recovery to pre-pandemic rates by week 38.

The model can be described as:

\[ y = \begin{cases} p1_x, & \text{if $x$ prior to lockdown}.\\ 0, & \text{if $x$ in lockdown}.\\ m(x-X), & \text{if $x$ in recovery}.\\ p2_x, & \text{otherwise} \end{cases} \] where: \(X\) is the last week in lockdown, \(x\) is the week,
\(p1_x\) is the 2020 value \(p2_x\) is the 2018-19 value



Note: the modelled line is a London-style recovery, not London’s exact 2020 trend. The model had to return to the region’s pre-pandemic level by the same date that London did, so London’s curve could not simply be applied to the other regions as pre-pandemic rates differed greatly.

The area between the modelled trend and the 2020 trend represent the admissions that did not happen but could have if other region’s recovery had matched that of London.

A total of 19,013 admissions for elective hip replacements could have occurred between mid-March and the end of 2020, if other regions had achieved the same recovery profile as London.



7 Notes and references

Notes:

COVID-19 lockdown dates: https://www.instituteforgovernment.org.uk/sites/default/files/timeline-lockdown-web.pdf

References:

[1] NHS England., 2021. Consultant-led Referral to Treatment Waiting Times Data 2020-21. Available at: https://www.england.nhs.uk/statistics/statistical-work-areas/rtt-waiting-times/rtt-data-2020-21/#Jan21 (Accessed 16th March 2021)

[2] Nikolova, S., Harrison, M. and Sutton, M., 2016. The impact of waiting time on health gains from surgery: Evidence from a national patient‐reported outcome dataset. Health economics, 25(8), pp.955-968.